Column Chromatography To Obtain Organic Cation Sorption

Subscriber access provided by La Trobe University Library

Article

Column Chromatography to Obtain Organic Cation Sorption Isotherms William C. Jolin, James Sullivan, Dharni Vasudevan, and Allison A MacKay Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b01733 • Publication Date (Web): 05 Jul 2016 Downloaded from http://pubs.acs.org on July 6, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30

Environmental Science & Technology

1

Column Chromatography to Obtain Organic Cation

2

Sorption Isotherms

3

William C. Jolin†, James Sullivanҗ, Dharni Vasudevanҗ, and Allison A. MacKay‡*

4

†

Department of Civil and Environmental Engineering, University of Connecticut, Storrs, CT

5

06269 җ

6 7

‡

Department of Chemistry, Bowdoin College, Brunswick, ME 04011

Department of Civil, Environmental and Geodetic Engineering, The Ohio State University,

8

9

Columbus, OH 43210

ABSTRACT

10

Column chromatography was evaluated as a method to obtain organic cation sorption

11

isotherms for environmental solids while using the peak skewness to identify the linear range of

12

the sorption isotherm. Custom packed HPLC columns and standard batch sorption techniques

13

were used to intercompare sorption isotherms and solid-water sorption coefficients (Kd) for four

14

organic

15

oxytetracycline) with two aluminosilicate clay minerals and one soil. A comparison of

16

Freundlich isotherm parameters revealed isotherm linearity or non-linearity was not significantly

17

different between column chromatography and traditional batch experiments. Importantly,

cations

(benzylamine,

2,4-dichlorobenzylamine,

phenyltrimethylammonium,

ACS Paragon Plus Environment

1


Page 2 of 30

18

skewness (a metric of eluting peak symmetry) analysis of eluting peaks can establish isotherm

19

linearity, thereby enabling a less labor intensive means to generate the extensive datasets of

20

linear Kd values required for the development of predictive sorption models. Our findings clearly

21

show that column chromatography can reproduce sorption measures from conventional batch

22

experiments with the benefit of lower labor-intensity, faster analysis times, and allow for

23

consistent sorption measures across laboratories with distinct chromatography instrumentation.

24 25

INTRODUCTION

26

Current sorption models, developed for neutral nonpolar compounds (e.g., EPISUITE1),

27

greatly underestimate sorption of ionic organic compounds because those models do not account

28

for electrostatic interactions between the sorbate and charged sorbents.2-6 Accurate sorption

29

models are necessary because a growing number of environmental contaminants of interest are

30

charged under environmental pH values, including surfactants, pesticides, antibiotics, and

31

pharmaceutical compounds.6-9 Recently, there has been encouraging progress in the development

32

of new predictive models for ionogenic compounds using polyparameter and computational

33

methods.3,4,10-13 In the case of cationic organic compounds, empirical sorption models using

34

molar volume and amine type (e.g., primary, secondary, or tertiary) show promise for predicting

35

sorption of non-heterocyclic amines of the form CxHyN to aluminosilicate clay minerals, organic

36

matter, and soils.3,4,10,14 Cationic amine compounds with more complex structures require the use

37

of corrective factors, derived from small sets of compounds sharing similar substructures (e.g.,

38

aromatic rings, -Cl). Sorption of structurally complex cations can be described using molecular

39

dynamics models that explicitly account for van der Waals and electrostatic energies of

40

interaction between the sorbate and the surface.12 However, advanced computational tools are


2

Page 3 of 30


41

specialized and not widely available to practitioners. In the absence of comprehensive sorption

42

models for organic cations, there is a need for a robust, time-efficient technique to obtain

43

sorption isotherms for ionic contaminants of interest and to generate extensive datasets of

44

sorption coefficients (Kd) for predictive model development.5,6,15,16

45

Sorption isotherms are obtained either by batch5,17-20 or column techniques.2,4,21 Batch isotherm

46

measurements consist of mixing the sorbent with an aqueous solution of the compound in a

47

closed reactor. Following sorptive equilibrium, the aqueous phase compound concentration is

48

measured and the sorbed concentration is often calculated from the difference between the initial

49

and final aqueous concentration with normalization to sorbent mass.22 Reactor preparation with

50

different initial solute concentrations allows the paired equilibrium sorbed (Cs, mmol kg-1) and

51

aqueous (Cw, mM) concentrations to be used to create an isotherm (a plot of Cs vs. Cw) for the

52

sorbent. The paired Cs and Cw values may be fit with the general Freundlich equation: ‫ܥ‬௦ = ‫ܭ‬௙ (‫ܥ‬ௐ )௡

(1)

53

where Kf (mmol kg-1 mM-n) is the Freundlich sorption coefficient and n is the Freundlich

54

exponent. A sorption coefficient can be obtained for any paired measure of Cs and Cw and is

55

defined as a single-point Kd (L kg-1). ‫ܭ‬ௗ =

‫ܥ‬௦ ‫ܥ‬௪

(2)

56

In cases where n ≈ 1, the Freundlich equation (eq 1) simplifies to the sorption coefficient.

57

Single-point Kd values that are constant with Cw are indicative of a linear sorption isotherm.

58

Obtaining sorption isotherms with batch experiments is straightforward, but can be time-

59

consuming and labor-intensive to include sufficient replicates and controls. These factors are

60

particularly important for sorption studies with large compound sets and for studies of ionogenic


3


Page 4 of 30

61

compounds that require assessment under many different experimental conditions (i.e., pH, ionic

62

strength, competing inorganic solutes).

63

Sorption coefficients obtained from column techniques have the advantage of preserving solids

64

in their native state.23,24 Column experiments to measure compound sorption are typically

65

employed to understand groundwater transport.25 Columns packed with the solid of interest are

66

flushed with synthetic groundwater at field flow rates. A step input of the compound of interest

67

is introduced to the column and compound breakthrough at the outlet of the column is tracked

68

over time. Sorption coefficients are derived from the retention time of the compound transported

69

through the column.26 Often, only one input concentration is utilized which can lead to

70

discrepancies between sorption coefficient measurements from column and batch experiments,27-

71

29

72

consistency between Kd values obtained from batch and column experiments, the experimental

73

conditions must yield Kd values in the linear range of the sorption isotherm (n = 1) and

74

breakthrough curves should be appropriately integrated.30 Further, the wide column diameters (>

75

several cm) typically employed to simulate groundwater transport can lead to disequilibrium

76

kinetics as diffusion times to sorption sites are long compared to advective transport through the

77

column.31-33 These factors, as well as the need for specialized equipment, have caused column

78

studies to be less favored than batch techniques34 for obtaining sorption isotherms.

particularly when isotherms for that sorbate-sorbent combination are nonlinear (n ≠ 1).30 For

79

Column chromatography, on the other hand, provides an effective means to overcome the

80

disadvantages of both batch and column techniques for obtaining sorption isotherms.35,36 Pulse

81

injections of a compound of known concentration are made to short (several cm length), narrow

82

diameter (10s mm) columns that are packed with a mixture of the sorbent of interest and an inert

83

non-sorptive solid (e.g., quartz35, SiC36). Compounds are then eluted with a mobile phase of


4

Page 5 of 30


84

known composition using a high-pressure pump. The inert solid restricts movement of small

85

sorbent particles within the column so that discrete sorbent particles contact the eluting phase,

86

thereby minimizing solid matrix diffusion effects and improving compound access to sorption

87

sites. As with large-scale columns, flowrates can still be varied to ensure equilibrium sorption

88

conditions and to further minimize kinetic mass transfer effects.35 Sorbent ‘dilution’ with the

89

inert sorbent allows the solid-to-water ratio of the sorbent to be varied to balance compound

90

separation from a non-retained tracer against dispersive effects which elongate compound

91

breakthrough curves. Shorter duration experiments are possible through inline detection of

92

compound breakthrough and automated calculation of retention times. Furthermore, the shape of

93

the eluting compound peak can provide information about isotherm non-linearity through

94

analysis of peak tailing or fronting.21,31,37-39 Transport of compounds under equilibrium sorption

95

conditions with non-linear isotherms will yield peak skewness values that differ from zero:

96

tailing (n < 1) gives positive values and fronting (n > 1) gives negative values.21,30 However,

97

skewness measures have not been validated as means of indicating isotherm non-linearity for

98

column chromatography.

99

To date, column chromatography to obtain sorption coefficients has been implemented by only

100

a limited number of research groups2,14,35,36,40, few of whom studied ionic sorbate

101

compounds.14,41 Droge et al used column chromatography measurements to collect an extensive

102

data set of organic cation sorption coefficients to organic matter, clay minerals and soils.3,4,10,14

103

Although equilibrium conditions for sorption coefficient measurements have been assessed

104

through the constancy of Kd values obtained under varied flow conditions37, direct comparisons

105

of isotherms obtained using column chromatography techniques with those from batch

106

techniques have only been performed for neutral compounds,40 while intercomparisons of


5


Page 6 of 30

107

column techniques across laboratories have not been undertaken. Ionic compounds are sensitive

108

to a number of system parameters, including equilibrium kinetics, eluent composition, and

109

sorbate chemistry thus, validation of column chromatography is needed to ensure sorption

110

coefficients from different collections means (batch to column, system to system) can be used in

111

developing predictive models.

112

The purpose of this study was to validate the column chromatography technique for measuring

113

sorption isotherms against conventional batch techniques. We examined four organic cations

114

with varied extents of isotherm linearity and nonlinearity (n =, >, or < 1) using different

115

environmentally relevant sorbents.

116

concentration ranges under which isotherm linearity could be assumed, we examined whether the

117

shape of the compound peak eluting from the column chromatography technique (measured as

118

skewness) showed regular trends with isotherm nonlinearity measures (i.e., n). Further,

119

transferability of the methodology and experimental findings were assessed across two labs with

120

distinctly different HPLC systems (U. Connecticut and Bowdoin College). Assessments of

121

method robustness are essential for broader adoption of column chromatography as a technique

122

for large-scale examination of sorption parameters so the Supplemental Information provides

123

guidance on the implementation of this method in other labs.

Because we were interested ultimately to discern

124 125

METHODS

126

Sorbents and chemicals

127

Texas Ca-montmorillonite (STx-1) and illite (IMt-1) were obtained from the Clay Minerals

128

Society. SiC was from Alfa Aesar. Iredell soil was collected and characterized previously.42

129

Sorbate compounds phenyltrimethylammonium (permanently charged), benzylamine (cationic,


6

Page 7 of 30


130

pKa: 9.33), 2,4-dichlorobenzylamine (9.15), and oxytetracycline (zwitterionic, pKa1: 3.27

131

(carboxylate, negative charge), pKa2: 7.41 (amine, positive charge)) were from Sigma Aldrich.

132

All other chemicals and reagents were ACS grade. Solutions were made with high purity 18.2

133

MΩ water (DI water) from a MilliQ system (Waters).

134

Batch sorption experiments

135

Batch sorption reactors were prepared with a solid-to-water ratio of 10 g/L for all sorbents.

136

Homoionic clays were created by washing (24 h) montmorillonite or illite three times with 1 M

137

sodium chloride, or 0.5 M calcium chloride, followed by three washes with DI water to remove

138

excess salt solution. Initial test compound concentrations (1 × 10-5 to 1 × 10-4 M) were chosen to

139

achieve an extent of equilibrium sorption between 5 and 95%, along with surface coverage

140

equivalent to less than 5% of the sorbent’s cation exchange capacity. For each test concentration,

141

paired Cw and Cs values were obtained in triplicate and included a control set with no sorbent

142

addition to infer initial compound concentration and to confirm the absence of other loss

143

processes. Sorbent-containing and sorbent-free reactors were prepared with background solutions

144

of either 5 mM CaCl2, 20 mM NaCl or DI water, mixed for 24 hours in the dark, centrifuged, and

145

filtered (0.45 µm PVTF) before analysis of the supernatant aqueous concentration. Aqueous

146

compound concentrations from the batch studies were determined using high pressure liquid

147

chromatography (Hewlett Packard HP 1050 outfitted with a C18 reverse phase column (Ultra

148

Aqueous, Restek) and a diode array detector). The system was operated with isocratic elution

149

with a mixture of solvents: (A) 20 mM phosphate buffer adjusted to pH 2.5 containing 4 mM

150

triethylammonium hydroxide and (B) acetonitrile. Mobile phase of (A: 90%, B: 10%) was used

151

for

phenyltrimethylammonium

and

benzylamine,

(A:

60%,

B:

40%)

for

2,4


7


Page 8 of 30

152

dichlorobenzylamine, and (A: 75%, B: 25%) for oxytetracycline. Absorbance wavelength of 205

153

nm was used for all compounds except oxytetracycline which was detected at 360 nm.

154

Column packing

155

Columns (30-mm length, 2.1-mm inner diameter, Restek #25118) were manually packed with

156

a mixture of silicon carbide (SiC) and sorbent material for the ‘sorbent-SiC’ columns and with

157

SiC for the ‘SiC-only’ columns. SiC-to-sorbent ratios (Table 1) were chosen so that the center of

158

mass of the breakthrough curves for the test compounds was at least 1.5 times greater than for a

159

non-retained tracer (NO3-) while minimizing peak spreading associated with extended compound

160

retention times. Also, each column was designed to have a sorbent-to-water ratio (Table 1) that

161

was within an order of magnitude of the 10 g/L used in the batch experiments. Dry SiC was

162

passed through a 200 mesh sieve (74 µm) and wet-filtered through a 0.75 µm glass fiber filter to

163

remove fines so particle sizes ranged from 0.75 to 74 µm. Montmorillonite was used as received.

164

Illite and Iredell soil were ground using a motor and pestle and then passed through 200 mesh

165

sieve. Fines were not removed from sorbent materials. Columns were packed without a column

166

vibrator using a procedure adapted from Bi et al.35 First, a mass of SiC and sorbent (typically 2 -

167

5 g of SiC, 10 - 50 mg sorbent) were combined together in a vial and vortexed for 2 minutes to

168

blend the powders uniformly. An aliquot (10 – 20 mg) of the combined solid phase mixture was

169

manually packed in a column using a spatula and tapping the column sides to limit mounding.

170

Periodically, the flat end of a thin rod was inserted into the column and used to lightly pack

171

down material so that particles were not clumped. The mass of the solid mixture (sorbent + SiC)

172

in the column was obtained from the difference in weights of the empty and packed sealed

173

columns. After packing, the sealed column was oriented vertically and attached to the HPLC

174

pump with flow directed vertically upward and discharging to waste. The column was slowly


8

Page 9 of 30


175

filled with DI water (10 µL/min) to remove all air. The flow rate was then increased gradually by

176

10 µL/min each hour to a final flowrate of 100 µL/min so that channeling within the packing

177

material was minimized. Solid-to-water ratios assumed the column void space to be completely

178

saturated. The difference in transport time of a non-retained tracer (NO3-) through the system

179

with a column in place and the tubing with no column attached was used to calculate the column

180

void space. The column void space was verified from the bulk density of the mixed column

181

packing using the known mass ratio and solid phase densities of the SiC and the sorbent material.

182

A comparative control column was packed entirely with SiC to verify that test compounds had

183

no sorptive interactions with this inert solid.

184

Table 1. Experimental Conditions used in the test columns. Column Name*

MMT-19

MMT-5

MMT-18B

ILL-53

IRE-23

Sorbent

Montmorillonite

Montmorillonite

Montmorillonite Illite

Iredell Soil

Void Space (µL)

45

45

45

63

48

Sorbent Mass (mg)

0.86

0.215

0.80

3.3

1.1

Sorbent-to-water ratio (g/L)

19

5

18

53

23

185

* Number adjacent to the column name refers to the sorbent-to-water ratio; B refers to column

186

packed at Bowdoin College; other columns were packed at U. Connecticut.

187

Column Operation

188

At U. Connecticut, packed columns were loaded into a standard HPLC system (Jasco PU-980

189

pump, AS-950 auto sampler, 40 µL injection loop, and MD-1510 multiwavelenth detector) with

190

tubing lengths minimized to reduce peak spreading. Aqueous phase solution chemistry was

191

fixed by the composition of column eluent solutions (Tab. 2). Aluminosilicate clay sorbents

192

were converted to homoionic form (Ca- and Na-montmorillonite; Na-illite) by flushing the


9


Page 10 of 30

193

respective column with air-equilibrated (pH 6 ± 0.3) 5 mM CaCl2 or 20 mM NaCl for 24 hrs

194

before compound injection. Compound sorption to Iredell soil was determined with DI water

195

adjusted to pH 5.2 (HCl). Only limited pre-flushing was performed for the Iredell soil column

196

since the use of DI water as an experimental eluent could alter the natural exchangeable ions

197

over time. (We note that experimentation with soil sorbents that is more extensive than our four

198

isotherm points would require an eluent solution that matches expected soil pore water

199

concentrations of natural exchange ions.) The effect of flow rate on sorption was evaluated by

200

varying flowrates between 50 and 200 µL min-1. Subsequently, an operating flowrate of 100 µL

201

min-1 was used for all experiments.

202

Compound or tracer solutions were introduced to the column using an injection volume of 40

203

µL and detected at the column outlet using wavelengths as indicated for the discrete sample

204

analyses from the batch experiments. The concentration of test compound injected into the

205

column was varied from 2.6 × 10-5 to 2.6 × 10-4 M to create sorption isotherms over a similar

206

range of sorbate concentrations as used in the batch experiments. Triplicate injections of each

207

concentration were made for both the SiC-only and the sorbent-SiC columns. Absorbance vs.

208

time data collected following each injection were exported directly to a MATLAB routine to

209

calculate the paired aqueous phase and sorbed concentrations from the eluted compound peak in

210

the breakthrough curve.

211

A similar column packing and operation procedure was used for the Bowdoin College HPLC

212

system (Agilent 1100 Series, Quaternary Pump, Diode Array Detector, and 100 µL injection

213

loop). The Bowdoin College HPLC system was more sensitive to the operating conditions for

214

obtaining sorption isotherms. The low eluent pumping rate (100 µL min-1), combined with a

215

slow syringe withdraw and eject speed (50 µL min-1) resulted in significant diffusional mixing of


10

Page 11 of 30


216

the compound with the mobile phase within the injection loop if it was partially filled. This

217

issue, not typically observed under normal analysis conditions (e.g., 1 mL min-1 flow rate), was

218

resolved by employing an injection volume of 100 µL to completely fill the injection loop.

219

Compounds with a single aromatic ring were detected at a wavelength of 210 nm (10 nm

220

bandwidth) because of high signal-to-noise ratios at lower detector wavelengths.

221

Mass balance

222

Mass balance assessments for batch experiments were undertaken for benzylamine sorption on

223

Ca-montmorillonite by adding CaCl2 salt to achieve a final concentration of 1 M and facilitate

224

competitive desorption. Samples were mixed for 24 hours in the dark, centrifuged, and filtered,

225

and an aliquot of the supernatant was removed for analysis by HPLC. Mass balance in the

226

column chromatography method was assessed by comparing the integrated compound peak areas

227

obtained with the ‘sorbent-SiC’ column with those from a ‘SiC-only’ column.

228

experimental approaches, mass balances within 5% were achieved, indicating the absence of any

229

other compound loss mechanisms for either sorption measurement techniques.

230

Isotherms

231

For both

Batch sorption experiments yielded equilibrium aqueous phase compound concentrations

232

directly.

Corresponding sorbed compound concentrations were calculated by difference as

233

outlined earlier.

234

Column sorption experiments required that sorption coefficients be calculated first using eluted

235

test compound peak characteristics, followed by extraction of the corresponding aqueous and

236

sorbed concentrations using a mass balance equation. For each compound mass injected, a Kd

237

value was calculated that corresponded to the center of mass (first moment33) or average travel

238

time of the compound peak eluted from the column33,36,42,43:


11


‫ܭ‬ௗ =

ܸ ܳ ∗ ൣ൫‫ݐ‬௖௠௣ௗ − ‫ݐ‬௧௥௔௖௘௥ ൯ − ൫‫ݐ‬௖௠௣ௗିௌ௜஼ − ‫ݐ‬௧௥௔௖௘௥ିௌ௜஼ ൯൧ = ݉ ݉

Page 12 of 30

(3)

239

where V (L) is the volume of eluent corresponding to the compound peak center of mass, m (mg)

240

is the column sorbent mass, and Q (L min-1) is the eluent flow rate. The bracketed term is the

241

effective travel time of the compound in contact with only the sorbent material in the ‘sorbent-

242

SiC’ column. Retention of the compound by the sorbent material was isolated from the overall

243

travel time through the ‘sorbent-SiC’ column (tcmpd – ttracer) by subtracting potential sorptive

244

interactions with SiC or other system components (tcmpd-SiC – ttracer-SiC). All times (ti, min) were

245

calculated by baseline-subtracting the pre-injection absorbance signal and then integrating the

246

corresponding peak to obtain the center-of-mass: tcmpd and tcmpd-SiC are the compound travel times

247

through the sorbent-SiC and SiC-only columns, respectively, and ttracer and ttracer-SiC are the

248

nitrate tracer travel times through sorbent-SiC and SiC-only columns, respectively. Sorption

249

coefficients obtained with Eq. 3 were used to obtain to theoretical effective aqueous and sorbed

250

concentrations for plotting as isotherm points36: ‫ܥ‬௪ = ‫ܥ‬௦ =

‫ܥ‬଴ ∗ ܸ௜ ‫ܭ‬ௗ ∗ ݉ + ܸ௜

‫ܥ‬଴ ∗ ܸ௜ − ‫ܥ‬௪ ∗ ܸ௜ ݉

(4) (5)

251

where C0 (mM) is the concentration of the injected test compound solution and Vi (L) is the

252

injection volume. Cw and Cs represent average concentration values for the particular sorbate-

253

sorbent conditions because the Kd value used directly in Eq. 4 and indirectly, via Cw, in Eq. 5 was

254

obtained from the center of mass as the pulse traveled through the column.42,43

255

Paired Cw and Cs values were used to construct isotherms. For the purposes of evaluating

256

isotherm linearity, single-point Kd values were calculated for every paired Cs and Cw value from

257

batch experiments, or obtained directly from column experiments (Eq. 3).20 Although single-


12

Page 13 of 30


258

point Kd values for potentially non-linear isotherms have received criticism when inter-

259

comparisons between compounds are not undertaken at the same Cw value, here, we use single

260

point Kd values to compare changes in slope within a single compound-sorbent isotherm. Kd

261

values were considered to be in the linear range (n = 1 approximation valid) if the lowest two or

262

more consecutive aqueous concentrations yielded Kd values that were not significantly different

263

from one another. Reported ‘linear’ Kd values (Kd_Linear) are the average of the individual, single-

264

point Kd values in the linear range of the isotherm.

265 266

Variations in symmetry of eluted compound peaks were examined for all Cw values used to construct isotherms by comparing peak skewness, S44: (‫ ݔ‬− ߤ)ଷ ܵ=෍ ߪଷ

(6)

267

where µ and σ are the mean and standard deviation of the individual points, x, of the

268

breakthrough curve.

269 270

RESULTS AND DISCUSSION

271

Equilibrium Assessment in Column Studies

272

Evidence for equilibrium sorption conditions within the columns was obtained by examining

273

the effect of flow rate on the sorption of benzylamine to a Na-montmorillonite column. It is

274

known that organic cations sorb quickly to soils and aluminosilicate clay minerals in batch

275

experiments, reaching equilibrium within eight hours;45,46 however, these timescales are much

276

greater than the 3- to 15-minute compound residence times observed in our columns. Sorption

277

disequilibrium in flow-through columns has been evidenced by lower sorption coefficients and

278

greater peak tailing as column flow rates are increased.39 For the three eluent flow rates tested at

279

U. Connecticut, benzylamine sorption to Na-montmorillonite (MMT-19, Table 1) showed no


13


Page 14 of 30

280

significant difference in Kd values: The 50 µL min-1 flow rate yielded a Kd value of 62 ± 1 L kg-

281

1

282

yielded a Kd value of 57 ± 5 L kg-1. Each of these values was not statistically different from the

283

Kd value of 61 ± 4 L kg-1 that was obtained from batch experiments, which had equilibrated for

284

24 hours (batch Cw values matched column conditions). Similarly, for the two eluent flow rates

285

tested in the Bowdoin College set up, phenyltrimethylammonium sorption to Ca-montmorillonite

286

(MMT-18B, Table 1) showed no significant difference in Kd values for flow rates of 50 µL min-1

287

(Kd = 126 ± 10 L/kg) and 100 µL min-1 (Kd = 124 ± 4 L/kg). These values were also close to the

288

Kd value of 132 ± 6 L/kg obtained in batch experiments. Greater tailing at the highest flow rate

289

(200 µL min-1) increased the measurement error (standard deviation of triplicate Kd

290

measurements) from 2% to 10%; therefore, 100 µL min-1 was used as the flow rate for all other

291

experiments. Previous researchers have also used 100 µL min-1 as an operating flowrate in

292

column chromatography to obtain sorption isotherms and sorption coefficients2,14; however, their

293

validation of equilibrium column conditions had not included higher column flow rates.

294

Isotherm Comparisons between Batch and Column Studies

, the 100 µL min-1 flow rate yielded a Kd value of 59 ± 1 L kg-1, and the 200 µL min-1 flow rate

295

The isotherms obtained from the column chromatography method reproduced the results from

296

conventional batch sorption experiments. Solid-to-water ratios of the sorbents in the column

297

were adjusted with SiC to match batch ratios so that direct comparisons at similar Cw values

298

could be made. Paired Cw and Cs values obtained from individual injections to the columns were

299

nearly identical to values obtained from batch reactors assembled under the same experimental

300

conditions (white v. black symbols, Fig. 1). In addition, paired Cw and Cs values obtained using

301

column chromatography at U. Connecticut and Bowdoin College were also nearly identical

302

(white squares without and with a plus, Fig 1A). Representative data shown in Fig. 1 includes


14

Page 15 of 30


303

compound-sorbent pairs chosen to validate the column chromatography method against batch

304

experiments under a range of sorption scenarios. First, we examined isotherms with differing n

305

values: Benzylamine sorption to Ca-montmorillonite (Fig. 1A) and phenyltrimethylammonium

306

sorption to Na-montmorillonite (Fig. 1B) exhibit Freundlich isotherms with (n > 1) because of

307

potential compound-compound interactions in the clay interlayers.20 Oxytetracycline sorption to

308

Na-montmorillonite (Fig. 1C) exhibits a traditional non-linear (n < 1) isotherm. Benzylamine

309

sorption to Iredell soil (Fig. 1D) and to illite (Fig. S1) are best described with linear isotherms.

310

Second, montmorillonite (Fig. 1A, B, C) and Iredell soil (Fig. 1D) both contain internal porosity

311

that could create mass transfer limitations with the shorter contact times for sorptive equilibrium

312

in column experiments, compared to batch reactors. The close match between column and batch

313

isotherms in each of these cases, as well as for additional compound-sorbent pairs (Fig. S1),

314

further confirms that equilibrium conditions established during column chromatography in two

315

different laboratory set-ups match those of conventional batch sorption experiments.

316


15


Page 16 of 30

317 318 319

Figure 1. Sorption isotherms obtained from column experiments match those obtained from

320

batch experiments for (A) benzylamine on Ca-montmorillonite, (B) phenyltrimethylammonium

321

on Na-montmorillonite, (C) oxytetracycline on Na-montmorillonite and (D) benzylamine on

322

Iredell soil. Black squares – batch data; white squares - column MMT-19 or IRE-23; white


16

Page 17 of 30


323

squares with black “+” – column MMT-18B; grey squares - column MMT-5. Where not visible,

324

error bars are smaller than the symbol size.

325 326

Agreement between column and batch methods was further supported by the observation that

327

isotherm parameter values (e.g., Kf and n) derived from regression fits to experimental batch

328

and column data were not statistically different (t-test, p > 0.05). Fitting the Freundlich equation

329

to paired Cw and Cs values gave similar Kf and n values for both methods for each of the eight

330

compound-sorbent pairs evaluated (Tab. 2). The one exception was oxytetracycline sorption to

331

Na-montmorillonite for which the three data points gave fits that were sensitive to small

332

variations in the sorbed concentration values. At low concentrations, where single point Kd

333

values were constant with Cw, Kd_Linear values were not statistically different between the two

334

methods either (Tab. 2). The standard deviation on isotherm parameters Kf and n or Kd_Linear

335

values were similar between the two methods (Tab. 2), although replicates of individual

336

concentration points showed greater reproducibility with the column chromatography method,

337

compared to replicate batch reactors.

338


17


Page 18 of 30

339

Table 2. Freundlich isotherm parameters (Kf and n exponent) and linear range sorption coefficients (Kd_Linear) obtained from batch and

340

column chromatography methods. Values in parentheses indicate data collected with Bowdoin College system. Compound

Solid

Background Solution

Batch Kf

Column Kf

Batch n

Column n

Batch Kd_Linear

Column Kd_Linear

Benzylamine

MMT

20 mM NaCl

98 ± 19§

105 ± 27

1.10 ± 0.05

1.14 ± 0.07

61 ± 4

59 ± 1

18 ± 6

27 ± 3

1.17 ± 0.03

1.10 ± 0.05

15 ± 2

17 ± 2

pH 6 Benzylamine

MMT

5 mM CaCl2

(29 ± 2)

(1.26 ± 0.02)

(12± 2.5)

pH 6 Benzylamine

IRE

DI pH 5.2

100 ± 60

115± 28

1.01 ± 0.41

1.01 ± 0.06

88 ± 8

96 ± 14

Benzylamine

ILL

20 mM NaCl

24 ± 3

27 ± 1

0.91 ± 0.03

0.93 ± 0.01

36 ± 3

38 ± 1

76800 ± 14700

65900 ± 4600

1.79 ± 0.16

1.74 ± 0.10

565 ± 62

556 ± 45

232 ± 32

201 ± 64

1.20 ± 0.05

1.12 ± 0.02

132 ± 6

131 ± 4

pH 6 Phenyltrimethylammonium

MMT

Phenyltrimethylammonium

MMT

20 mM NaCl pH 6 5 mM CaCl2

(194 ± 24)

(1.11 ± 0.01)

(133± 16)

pH 6


18

Page 19 of 30

341


Oxytetracycline

MMT

20 mM NaCl pH 6.5

27 ± 9

19 ± 1

0.44 ± 0.08

0.34 ± 0.04

449 ± 10

415 ± 26

2,4dichlorobenzylami ne

MMT

20 mM NaCl

108 ± 21

120 ± 10

1.08 ± 0.04

1.10 ± 0.04

67 ± 4

72 ± 3

§

pH 6

represents average value and standard deviation based on triplicate analyses


19


342

Page 20 of 30

Concentration ranges amenable to column chromatography

343

The range of concentrations over which isotherms can be obtained by column

344

chromatography is set by physical limitations of the system detector. The lower concentration

345

bound of an isotherm is set by the limit of detection of the instrument which can be as low as 5 ×

346

10-7 M for modern absorbance detection. Low concentrations are also of concern for compounds

347

that exhibit high extents of sorption and consequently, long retention times. Long retention

348

times result in peak spreading that can lower compound absorbance to background levels. Such

349

situations can be addressed by reducing the ratio of sorbent to inert packing material in a column.

350

We evaluated isotherm consistency across columns with different sorbent ratios using

351

phenyltrimethylammonium, the most sorptive compound of the test compounds examined here.

352

Isotherm points obtained on a column with an effective solid-to-water ratio of 5 g/L of Na-

353

montmorillonite (MMT-5, grey squares, Fig. 1B) compared well to those obtained with a Na-

354

montmorillonite column packed with four times the mass of sorbent of 19 g/L (MMT-19, white

355

squares, Fig. 1B). Individual data points for the MMT-19 column showed greater measurement

356

errors as a result of greater peak spreading associated with the longer compound retention time

357

(MMT-19: 8-15 min, MMT-5: 3–6 min). Thus, adjustments of the sorbent-to-SiC ratio in the

358

chromatography columns will yield reproducible isotherms while enhancing peak detection via

359

shorter column retention times. Other means to improve detection of highly retained compounds

360

are discussed in the Supporting Information.

361

Our primary interest in this study was to obtain sorption measurements under conditions of

362

very low sorbent coverage (< 2% of cation exchange sites) that were within the linear range of

363

the isotherm.

364

benzylamine sorption to Na-montmorillonite up to 100% cation exchange site coverage. This

For completeness, we used the column chromatography method to examine


20

Page 21 of 30


365

compound, like other cationic aromatic amines, displays an ‘S-shaped’ non-linear isotherm that

366

curves away from the x-axis at low concentrations due to sorbate-sorbate cation-π interactions on

367

the surface, followed by curvature toward the x-axis as the sorbed concentration approaches the

368

cation exchange capacity.20 The complex S-shape isotherm for benzylamine sorption to Na-

369

montmorillonite previously observed in batch studies20 was successfully reproduced by our

370

column chromatography method (Fig. 2). Quantitative comparisons with previous batch studies

371

are not shown in Fig. 2 because of differences in clay mineral exchange ions (e.g. hetero vs.

372

homoionic clay minerals) between the two studies. Nevertheless, the data in Fig. 2 demonstrates

373

the ability of the column method to capture isotherm points up to full surface coverage without

374

reaching detector saturation at the high concentrations injected.

375 376

Figure 2. Isotherm ranging from 0.1 to 100 percent exchange site coverage for benzylamine on

377

Na-montmorillonite using column chromatography displays complex “S” shape. Dashed line

378

indicates cation exchange capacity of montmorillonite.

379 380

Breakthrough curve skewness as an indicator of non-linearity


21


Page 22 of 30

381

We explored the use of compound peak symmetry measures to bound the range of Cw values

382

over which isotherm linearity can be assumed (n = 1 valid). Kd_Linear values are often required for

383

calibrating predictive models.3,12,16 Such values are typically obtained by performing batch

384

experiments to collect isotherms and comparing single-point Kd values (Eq. 2) across these

385

extensive datasets. The labor-intensity of such an approach could be avoided if it were possible

386

to assess whether a Cw and Cs pair falls within the linear isotherm range, independent of the full

387

isotherm dataset. Accordingly, we investigated whether peak skewness measures might provide

388

such insights. Under transport conditions with local sorptive equilibrium, peak symmetry is non-

389

uniform when the Freundlich parameter deviates from n = 1.21,30,33,37-39 We postulated that such

390

a change in peak skewness might be observable across the set of points constituting an isotherm.

391 392

For our column operating conditions, the most appropriate comparative measure of peak symmetry was the fractional change in skewness, %∆S: %∆ܵ =

ܵ݇݁‫ݏݏ݁݊ݓ‬௦௢௥௕௘௡௧ିௌ௜஼ − ܵ݇݁‫ݏݏ݁݊ݓ‬ௌ௜஼ × 100%. ܵ݇݁‫ݏݏ݁݊ݓ‬ௌ௜஼

(7)

393

where Skewnessi is the calculated skewness for the compound peak on the respective ‘sorbent-

394

SiC’ and ‘SiC-only’ columns. A difference measure was implemented because peak asymmetry

395

was observed for eluted peaks of both the tracer and test compounds on the non-sorptive ‘SiC-

396

only’ column. This asymmetry likely originated from the injection volumes being greater than

397

1/6 of the column volume47 and skewness values were the same for all compounds on the ‘SiC-

398

only’ column. Skewness measures for the non-retained tracer were the same for the system: (i)

399

without a column attached, (ii) with the ‘SiC-only’ column, and (iii) with the ‘sorbent-SiC’

400

columns, indicating column packing to be homogenous and consistent across the different

401

column preparations (Table S1). Further, for each compound, the same peak skewness was

402

calculated at all injected concentrations less than 2 mM on the ‘SiC-only’ column. The only


22

Page 23 of 30


403

exceptions were the three highest concentrations (6 mM to 20 mM) used to obtain the

404

benzylamine isotherm (Fig. 2). Thus, we concluded that variations in %∆S across the isotherm

405

for a given test compound could be ascribed to variations in equilibrium Kd and that the %∆S

406

parameter could be used to assess isotherm non-linearity.

407

The fractional change in peak skewness appears to be a robust indicator of single-point Kd

408

linearity. Single-point Kd values were normalized to the average Kd-Linear to assess the fractional

409

deviation from linearity: %∆‫ܭ‬ௗ =

‫ܭ‬ௗ,௣௢௜௡௧ − ‫ܭ‬ௗ,௟௜௡௘௔௥ × 100%. ‫ܭ‬ௗ,௟௜௡௘௔௥

(8)

410

Only paired Cw and Cs values that had a small fractional change in skewness less than |± 4%|

411

were from the linear range of the sorption isotherm (i.e., %∆Kd ~ 0) (white symbols, Fig. 3).

412

Although the absolute skewness values were distinct for measurements made at U. Connecticut

413

(1.5) and Bowdoin (2.5), a %∆S less than an absolute value of 4% for experiments conducted at

414

either location indicated that sorption was in the linear range of the isotherm (Fig. 3A). For high

415

concentration points of compound-sorbent pairs from isotherms with Freundlich exponents n > 1

416

(i.e., %∆Kd > 1), %∆S exhibited negative values and fell in the upper left quadrant of Fig. 3B.

417

On the other hand, high concentration points for isotherms with n < 1 (i.e., %∆Kd < 1), %∆S

418

exhibited positive values and fell in the lower right quadrant (Fig. 3B). From this analysis it is

419

clear that %∆S for a given concentration injection that is more positive, or more negative, than

420

4% is indicative of sorption in the non-linear range of the sorption isotherm. Therefore, the

421

relative change in skewness can be used to bound the linear range of a sorption isotherm.

422

Outlook

423

The indication of linear range sorption coefficients for non-heterocyclic compounds,

424

coupled with the already reduced labor-intensity compared to traditional batch experiments,


23


Page 24 of 30

425

makes column chromatography an efficient, robust tool to collect sorption coefficients of organic

426

cations. We revisited column chromatography as a technique to obtain sorption coefficients for

427

environmental solids because it has been underutilized in sorption studies. Yet, the labor-saving

428

benefits of measuring sorption coefficients and isotherms via column chromatography are

429

advantageous to studies that require sorbate characterization for large sets of compounds, such as

430

by regulatory agencies. This this end, we detail in the Supplemental Information a strategy for

431

successful implementation of column chromatography by other experimentalists, based upon our

432

efforts to transfer the technique between labs. Finally, the availability of a quantitative

433

parameter (%∆S) to bound isotherm linearity will be beneficial in the development of descriptive

434

sorption models for ionizable compounds with a limited concentration range of isotherm

435

linearity.

436 437 438

439 440

Figure 3. Relationship between %∆Kd and %∆S for (A) paired Cs and Cw values near the

441

linear range of the sorption isotherm and (B) all sorbate-sorbent pairs in linear and non-linear


24

Page 25 of 30


442

range shows that changes in skewness of less than |± 4%| were associated with single-point Kd

443

values that varied by 5% or less from independently assessed ‘linear’ Kd values (white circles, U.

444

Connecticut; grey circles, Bowdoin College).

445

ASSOCIATED CONTENT

446

Supporting Information. A guide to implementing column chromatography onto new systems,

447

a discussion on possible improvement in peak detections, comparison of peak skewness as it

448

relates to column homogeneity, and additional isotherms comparing column chromatography

449

with batch experiments are available in the Supplemental Information. The Matlab code is

450

available on request. This material is available free of charge via the Internet at

451

http://pubs.acs.org.

452

AUTHOR INFORMATION

453

Corresponding Author

454

*[email protected]

455

Author Contributions

456

The manuscript was written through contributions of all authors. All authors have given approval

457

to the final version of the manuscript.

458

ACKNOWLEDGMENT

459

The authors acknowledge funding from the NSF through CHE Grants #1404998 and #1404459.

460

We thank the three anonymous reviewers for their constructive comments which helped us

461

improve the manuscript.

462

Table of Contents Graphic


25


Page 26 of 30

463 464

465 466

1. U.S. Environmental Protection Agency Estimation Programs Interface Suite for Microsoft Windows (EPISUITE), Washington, DC, 2012; Version 4.11.

467 468

2. Bi, E.; Schmidt, T.C.; Haderlein, S.B. Environmental factors influencing sorption of heterocyclic aromatic compounds to soil. Environ. Sci. Technol. 2007, 41 (9), 3172-3178.

469 470 471

3. Droge, S. and Goss, K. Development and Evaluation of a New Sorption Model for Organic Cations in Soil: Contributions from Organic Matter and Clay Minerals. Environ. Sci. Technol. 2013, 47 (24), 14233-14241.

472 473 474

4. Droge, S. and Goss, K. Ion-Exchange Affinity of Organic Cations to Natural Organic Matter: Influence of Amine Type and Nonionic Interactions at Two Different pHs. Environ. Sci. Technol. 2013, 47 798-806.

475 476 477

5. Williams, M.; Ong, P.L.; William, D.B.; Kookana, R.S. Estimating the Sorption of Pharmaceuticals Based on their Pharmacological Distribution. Environ. Toxicol. Chem. 2009, 28 (12), 2572-2579.

478 479

6. Tolls, J. Sorption of Veterinary Pharmaceuticals in Soils: A Review. Environ. Sci. Technol. 2001, 35 (17), 3397-3406.

480 481 482 483

7. Kolpin, D.W.; Furlong, E.T.; Meyer, M.T.; Thurman, E.M.; Zaugg, S.D.; Barber, L.B.; Buxton, H.T. Pharmaceuticals, Hormones, and Other Organic Wastewater Contaminants in U.S. Streams, 1999−2000, A National Reconnaissance. Environ. Sci. Technol. 2002, 36 (6), 12021211.


26

Page 27 of 30


484 485

8. Sim, W.; Lee, J.; Oh, J. Occurrence and fate of pharmaceuticals in wastewater treatment plants and rivers in Korea. Environ. Poll. 2010, 158 (5), 1938-1947.

486 487

9. Zuccato, E.; Castiglioni, S.; Bagnati, R.; Chiabrando, C.; Grassi, P.; Fanelli, R. Illicit drugs, a novel group of environmental contaminants. Water Res. 2008, 42 (4), 961-968.

488 489 490

10. Droge, S. and Goss, K. Sorption of Organic Cations to Phyllosilicate Clay Minerals: CECNormalization, Salt Dependency, and the Role of Electrostatic and Hydrophobic Effects. Environ. Sci. Technol. 2013, 47 (24), 14224-14232.

491 492 493

11. Jadbabaei, N. and Zhang, H. Sorption Mechanism and Predictive Models for Removal of Cationic Organic Contaminants by Cation Exchange Resins. Environ. Sci. Technol. 2014, 48 (24), 14572-14581.

494 495 496

12. Samaraweera, M.; Jolin, W.; Vasudevan, D.; MacKay, A.A.; Gascon, J.A. Atomistic Prediction of Sorption Free Energies of Cationic Aromatic Amines on Montmorillonite: A Linear Interaction Energy Method. Environ. Sci. Technol. Lett. 2014, 1 (6), 284-289.

497 498 499

13. Zhang, H.; Shields, A.J.; Jadbabaei, N.; Nelson, M.; Pan, B.; Suri, R.P.S. Understanding and Modeling Removal of Anionic Organic Contaminants (AOCs) by Anion Exchange Resins. Environ. Sci. Technol. 2014, 48 (13), 7494-7502.

500 501

14. Droge, S. and Goss, K. Effect of Sodium and Calcium Cations on the Ion-Exchange Affinity of Organic Cation for Soil Organic Matter. Environ. Sci. Technol. 2012, 46 (11), 5894-5901.

502 503 504

15. MacKay, A. and Vasudevan, D. Polyfunctional Ionogenic Compound Sorption: Challenges and New Approaches To Advance Predictive Models. Environ. Sci. Technol. 2012, 46 92099223.

505 506

16. Endo, S. and Goss, K. Applications of Polyparameter Linear Free Energy Relationships in Environmental Chemistry. Environ. Sci. Technol. 2014, 48 (21), 12477-12491.

507 508

17. Chiou, C.T.; Peters, L.J.; Freed, V.H. A physical concept of soil-water equilibria for nonionic organic compounds. Science 1979, 206 (4420), 831-832.

509 510

18. Chiou, C.T.; Porter, P.E.; Schmedding, D.W. Partition equilibriums of nonionic organic compounds between soil organic matter and water. Environ. Sci. Technol. 1983, 17 (4), 227-231.

511 512 513

19. Carrasquillo, A.; Bruland, G.; MacKay, A.; Vasudevan, D. Sorption of Ciprofloxacin and Oxytetracycline Zwitterions to Soils and Soil Minerals: Influence of Compound Structure. Environ. Sci. Technol. 2008, 42 7634-7642.

514 515 516

20. Vasudevan, D.; Arey, T.; Dickstein, D.; Newman, M.; Zhang, T.; Kinnear, H.; Bader, M. Nonlinearity of Cationic Aromatic Amine Sorption to Alminosilicates and Soils: Role of Intermolecular Cation-pi Interactions. Environ. Sci. Technol. 2013, 47 (24), 14119-14127.


27


Page 28 of 30

517 518 519

21. Spurlock, F.C.; Huang, K.; van Genuchten, M.T. Isotherm nonlinearity and nonequilibrium sorption effects on transport of fenuron and monuron in soil columns. Environ. Sci. Technol. 1995, 29 (4), 1000-1007.

520 521

22. Weber, W.J. and Usinowics, P. Adsorption from aqueous solution. American Chemical Society: 1968.

522 523 524

23. Holmén, B.A. and Gschwend, P.M. Estimating sorption rates of hydrophobic organic compounds in iron oxide-and aluminosilicate clay-coated aquifer sands. Environ. Sci. Technol. 1996, 31 (1), 105-113.

525 526 527

24. Roberts, P.V.; Goltz, M.N.; Mackay, D.M. A natural gradient experiment on solute transport in a sand aquifer: 3. Retardation estimates and mass balances for organic solutes. Water Resour. Res. 1986, 22 (13), 2047-2058.

528 529 530

25. Van Genuchten, M.T.; Wierenga, P.; Klute, A. Solute dispersion coefficients and retardation factors. In Methods of Soil Analysis: Part 1—Physical and Mineralogical Methods, American Society of Agronomy, Inc.: Madison, WI, 1986; pp. 1025-1054.

531 532

26. Relyea, J.F. Theoretical and experimental considerations for the use of the column method for determining retardation factors. Radioact. Waste Manage. Nucl. Fuel Cycle 1982, 151-166.

533 534 535

27. Porro, I.; Newman, M.E.; Dunnivant, F.M. Comparison of batch and column methods for determining strontium distribution coefficients for unsaturated transport in basalt. Environ. Sci. Technol. 2000, 34 (9), 1679-1686.

536 537 538

28. Fouquet, Y.; von Stackelberg, U.; Charlou, J.L.; Donval, J.P.; Foucher, J.P.; Erzinger, J.; Herzig, P.; Mühe, R.; Wiedicke, M.; Soakai, S. Hydrothermal activity in the Lau back-arc basin: Sulfides and water chemistry. Geology 1991, 19 (4), 303-306.

539 540

29. Wang, T.; Li, M.; Teng, S. Bridging the gap between batch and column experiments: A case study of Cs adsorption on granite. J. Hazard. Mater. 2009, 161 (1), 409-415.

541 542 543

30. Bürgisser, C.S.; Cernik, M.; Borkovec, M.; Sticher, H. Determination of nonlinear adsorption isotherms from column experiments: An alternative to batch studies. Environ. Sci. Technol. 1993, 27 (5), 943-948.

544 545 546

31. Maclntyre, W.G.; Stauffer, T.B.; Antworth, C.P. A comparison of sorption coefficients determined by batch, column, and box methods on a low organic carbon aquifer material. Groundwater 1991, 29 (6), 908-913.

547 548 549

32. Allen, H.E.; Chen, Y.; Li, Y.; Huang, C.; Sanders, P.F. Soil partition coefficients for Cd by column desorption and comparison to batch adsorption measurements. Environ. Sci. Technol. 1995, 29 (8), 1887-1891.


28

Page 29 of 30


550 551

33. Valocchi, A.J. Validity of the local equilibrium assumption for modeling sorbing solute transport through homogeneous soils. Water Resour. Res. 1985, 21 (6), 808-820.

552 553

34. OECD Test No. 106: Adsorption -- Desorption Using a Batch Equilibrium Method. Organization for Economic Co-operation and Development: 2000.

554 555

35. Bi, E.; Schmidt, T.C.; Haderlein, S.B. Practical issues relating to soil column chromatography for sorption parameter determination. Chemosphere 2010, 80 (7), 787-793.

556 557 558 559

36. Schenzel, J.; Goss, K.; Schwarzenbach, R.P.; Bucheli, T.D.; Droge, S.T. Experimentally Determined Soil Organic Matter–Water Sorption Coefficients for Different Classes of Natural Toxins and Comparison with Estimated Numbers. Environ. Sci. Technol. 2012, 46 (11), 61186126.

560 561

37. Bi, E.; Zhang, L.; Schmidt, T.C.; Haderlein, S.B. Simulation of nonlinear sorption of Nheterocyclic organic contaminates in soil columns. J. Contam. Hydrol. 2009, 107 (1), 58-65.

562 563 564

38. Fesch, C.; Simon, W.; Haderlein, S.B.; Reichert, P.; Schwarzenbach, R.P. Nonlinear sorption and nonequilibrium solute transport in aggregated porous media: Experiments, process identification and modeling. J. Contam. Hydrol. 1998, 31 (3), 373-407.

565 566 567

39. Altfelder, S.; Streck, T.; Maraqa, M.A.; Voice, T.C. Nonequilibrium sorption of dimethylphthalate—Compatibility of batch and column techniques. Soil Sci. Soc. Am. J. 2001, 65 (1), 102-111.

568 569

40. Bronner, G. and Goss, K. Sorption of organic chemicals to soil organic matter: influence of soil variability and pH dependence. Environ. Sci. Technol. 2010, 45 (4), 1307-1312.

570 571

41. Tülp, H.C.; Fenner, K.; Schwarzenbach, R.P.; Goss, K. pH-dependent sorption of acidic organic chemicals to soil organic matter. Environ. Sci. Technol. 2009, 43 (24), 9189-9195.

572

42. Jonsson, J.A. Chromatographic theory and basic principles. CRC Press: 1987.

573 574

43. Helfferich, F. Travel of molecules and disturbances in chromatographic columns: A paradox and its resolution. J. Chem. Educ. 1964, 41 (8), 410.

575 576

44. MacGillivray, H. Skewness and asymmetry: measures and orderings. Ann. of Stat. 1986, 9941011.

577 578

45. Jones, A.D.; Bruland, G.L.; Agrawal, S.G.; Vasudevan, D. Factors influencing the sorption of oxytetracycline to soils. Environ. Toxicol. Chem. 2005, 24 (4), 761-770.

579 580

46. Brownawell, B.; Chen, H.; Collier, J.; Westall, J. Adsorption of Organic Cations to Natural Materials. Environ. Sci. Technol. 1990, 24 (8), 1234-1241.

581

47. Kromidas, S. More practical problem solving in HPLC. John Wiley & Sons: 2008.


29


Page 30 of 30

582


30

Column Chromatography To Obtain Organic Cation Sorption

Recommend Documents